neurocore.llm.provider.LLMProvider
==================================

.. py:class:: neurocore.llm.provider.LLMProvider

   Bases: :py:obj:`Protocol`


   Protocol for LLM provider implementations.

   Skills that need LLM access declare:
       skill_meta = SkillMeta(..., requires_llm=True)

   NeuroCore injects self.llm during skill init when requires_llm=True.


   .. py:property:: provider_name
      :type: str



   .. py:property:: model
      :type: str



   .. py:method:: complete(messages: list[LLMMessage], *, max_tokens: int = 8192, temperature: float = 1.0, system: str | None = None, **kwargs: Any) -> LLMResponse
      :async:



   .. py:method:: stream(messages: list[LLMMessage], *, max_tokens: int = 8192, temperature: float = 1.0, system: str | None = None, **kwargs: Any) -> collections.abc.AsyncIterator[str]
      :async: